Academic

Tillers on Universal Principles of Evidential Inference

By Lawrence Solum

Peter Tillers (Cardozo Law School) has posted Are There Universal Principles or Forms of Evidential Inference? on SSRN.  Here is the abstract:

Although interest in evidential inference is not new - interest in the topic reaches back into antiquity - during the last two or three decades there has been a veritable explosion of scholarship and research about evidential inference. Furthermore, evidential inference (or "factual inference") is now an important topic in virtually every field of scholarship and in virtually every kind of "knowledge industry." Although the models of inference generated in this latest wave of scholarship and research are varied, one thread does run through many of the new models. Many contemporary accounts emphasize the multistage nature of evidential inference; it is now very often argued or assumed that evidential inference is best viewed as a network or web of inferences Although the proponents of such models of evidential inference often have important disagreements about the properties or structure of multistage evidential inference, it is fair to say that such models generally rest on the compound proposition that real-world evidential inference usually or always consists of "atoms" that are linked together (in some way) by "generalizations." If a model of this sort is a valid representation of evidential inference, the question may arise - the question has arisen - whether a model of this sort is or is not "universal." The answer is that this question is unanswerable. What can be said is that when human beings (or other agents) configure problems of evidence in a certain way, inference networks (of some sort) are inevitable and describe the structure of the problem at hand but that when problems of evidence are perceived (only) in certain other ways, representations of inference as webs of factual hypotheses connected by generalizations are of little or no use - simply because in some situations such web-like patterns of reasoning do not address the problem at hand. For example, sometimes the question is not how strongly some evidentiary phenomenon supports some hypothesis but, rather, what sort of complex of hypotheses or conceptual constructs most persuasively explains some set of phenomena. This latter sort of problem sometimes requires a constructive and imaginative conceptual activity that does not much resemble an inference network or its ingredients.

I always learn from Tiller.

Full post as published by Legal Theory Blog on January 14, 2008 (boomark / email).